The Moddel cell and the thermodynamic ledger

Why the library suggests it

Garret Moddel reports the device producing electrical power with no apparent input, at a power density of about 70 watts per square metre, with the trend measured across thousands of devices in dozens of batches and nine candidate measurement artefacts investigated (Zero-Point Energy: Capturing Evanescence, 2022). The patent gives the fabrication recipe in micrometre-scale cells of chromium, silicon dioxide and aluminium and puts a number on a single device — about 0.66 nanowatts up to a terahertz cut-off (Quantum noise power devices, 2020). The same author's review sorts every published extraction scheme into three families and rules two of them out on thermodynamic grounds, leaving the geometry-dependent route the cell actually uses — and reports that his own gas-flow test measured radiation about a thousand times weaker than predicted (Extraction of Zero-Point Energy from the Vacuum, 2019). The thermodynamic permission slip is older still: Daniel Cole and Harold Puthoff worked the entropy through both a reversible squeeze and an irreversible one and found the books balance (Extracting energy and heat from the vacuum, 1993), and the gas-through-cavities architecture is set out with materials and a power budget in Quantum vacuum energy extraction (2005).

The experiment or build

Fabricate the cell from the published recipe in a second laboratory. Run the full artefact list — thermoelectric offsets, rectified ambient radio-frequency, contact potentials, drift — and publish it whatever the answer. Then do the control the field most needs: identical diodes with and without the cavity, on the same wafer, measured in the same session. The settling measurement is output power per unit area from cavity-bearing and cavity-free devices on one wafer, with the difference surviving the artefact list. Keep a clean comparison alongside it: Paul Thibado's graphene harvester rectifies the thermal ripple of a freestanding sheet and reached 40 picowatts at one diode, with the accounting explicitly at a single temperature (Fluctuation-induced current from freestanding graphene, 2020; Spectrum Analysis of Thermally Driven Curvature Inversion in Strained Graphene Ripples, 2025). That device draws on thermal, not vacuum, fluctuations, and keeping the two apart is what makes either claim assessable.

Moddel's cell is the most advanced hardware claim on that board, and the right way to hold a hardware claim is as a ledger. Here it is: the claim on one side, every artefact that has been tested and how it was excluded on the other, and the one control the field most needs.

The Moddel cell, and the ledger that would settle it

A tunnelling diode about two nanometres across with a Casimir cavity on top, reported at about seventy watts per square metre with no apparent input. This is the ledger: the claim on one side, every artefact that has been tested and how it was excluded on the other, and the one control the field most needs. Filter the rows by what they are, and by which apparatus they belong to.

The claim, as its author states it

Reported power density

70 W/m²

For comparison, a solar cell over the same area

210 W/m²

Insulator the electrons tunnel through

2 nm

One patented device, up to a terahertz cut-off

6.6E-10 W

Candidate measurement artefacts investigated by the author

9

Read the claim at a size you can picture

Device area, in square centimetres

1.0

Output at the reported density

7E-3 W

Single patented devices needed for one watt

1.52E9

• Measured evidence — 5
• Artefact excluded — 5
• Argued against, not ruled out — 1
• Left open by the authors — 2

Filter by what the row isEvery kind of rowMeasured evidenceArtefact excludedArgued against, not ruled outLeft open by the authorsFilter by apparatusEvery apparatusThe tunnelling diode with a Casimir cavityThe gas-flow experimentThe thermodynamic route analysisThe clean comparison beside it

Clear both filters

Measured evidence: 5. Artefact excluded: 5. Argued against, not ruled out: 1. Left open by the authors: 2

The row	What it is	Which apparatus
Short-circuit output rises as the Casimir cavity is made thinner, for cavities filled with PMMA and with silicon dioxide alike.This is the trend the author identifies as the signature to look for, because a thinner cavity suppresses more of the long-wavelength modes. It is a dependence on geometry, which is what separates this claim from a thermal one.The paper	Measured evidence	The tunnelling diode with a Casimir cavity
The trends are reported across many thousands of devices produced in dozens of batches.A count of devices is not a replication, and the paper does not present it as one. It is what makes a systematic fabrication artefact less likely, and it is reported here as exactly that.The paper	Measured evidence	The tunnelling diode with a Casimir cavity
Nine candidate measurement artefacts were investigated by the author, and none was able to explain the observed results.The enumeration is in that paper's own appendix, which this site does not reproduce. The count is reported as a count.The paper	Artefact excluded	The tunnelling diode with a Casimir cavity
A Casimir resistor through a high-frequency line into a diode with a terahertz cut-off gives 0.66 nanowatts per device at a matched load.The patent also gives the recipe: micrometre-scale cells of chromium, silicon dioxide and aluminium, wired into series and parallel arrays.The paper	Measured evidence	The tunnelling diode with a Casimir cavity
The Joule-Thomson effect, in the gas-flow experiment.Excluded: the lock-in phase was inverted, which indicates cooling rather than the heating the artefact would produce.The paper	Artefact excluded	The gas-flow experiment
Turbulence in the flowing gas, in the gas-flow experiment.Excluded: the Reynolds number sits far below the turbulent threshold at flows under two centimetres per second.The paper	Artefact excluded	The gas-flow experiment
Absorption or adsorption at the membrane, in the gas-flow experiment.Excluded: the signal did not scale with membrane surface area, which it would have to if the surface were producing it.The paper	Artefact excluded	The gas-flow experiment
Frictional heating of the gas, in the gas-flow experiment.Argued against but not ruled out: the ordering across gases comes out wrong, and the dependence on flow is linear where friction would make it quadratic. The authors say so themselves.The paper	Argued against, not ruled out	The gas-flow experiment
The measured power stayed below one microwatt for every gas and sample, about three orders of magnitude under the prediction.The authors call the result tantalising but inconclusive, and name what they would do next: nanopores of consistent size and shape, the radiation imaged to find its source, and its spectrum measured.The paper	Left open by the authors	The gas-flow experiment
Rectifying the zero-point field with a diode, or with an antenna feeding a diode.Excluded as a route, by the same author's own review: the field is in true equilibrium, so detailed balance binds each step separately and no net power can be drawn that way.The paper	Artefact excluded	The thermodynamic route analysis
Extraction from Casimir cavities is consistent with thermodynamics in principle.Cole and Puthoff worked the entropy through both a reversible squeeze and an irreversible one, and the books balance. The permission slip is old and it is real; what it permits is a device, not a conclusion.The paper	Measured evidence	The thermodynamic route analysis
Where the energy ultimately comes from.Left open by the author in his own words. That is the right way to leave it, and it is the question the settling measurement below would begin to answer.The paper	Left open by the authors	The thermodynamic route analysis
A freestanding graphene sheet rectified into a diode circuit reached forty picowatts at one diode.Kept beside the cell as the clean comparison, not as support: that device draws on thermal fluctuations at a single temperature, and its authors keep the accounting that way.The paper	Measured evidence	The clean comparison beside it

The measurement that would settle it

Output power per unit area from cavity-bearing and cavity-free devices on one wafer, measured in the same session by a second laboratory, with the difference surviving the artefact list. One wafer, two populations, one session: that is the control the field most needs and the one nobody has published.

Keep the two claims apart, because keeping them apart is what makes either one assessable. Thibado's graphene harvester rectifies the thermal ripple of a freestanding sheet and reached forty picowatts at one diode, with the accounting explicitly at a single temperature. That device draws on thermal fluctuations. The cell above is a claim about geometry.

On the bench now. Scope: a ledger of what has been measured and what has been excluded, with every row sourced. It is not a verdict, and it does not need to be one — the settling measurement is named above and somebody can go and take it.